Coding Device, Decoding Device, and Method and Program Thereof

This application is a continuation of and claims the benefit of priority under 35 U.S.C. §120 to U.S. application Ser. No. 18/743,662 filed Jun. 14, 2024, which is a continuation of U.S. application Ser. No. 18/195,015 filed May 9, 2023 (now U.S. Pat. No. 12,051,430, issued Jul. 30, 2024), which is a continuation of U.S. application Ser. No. 17/370,060 filed Jul. 8, 2021 (now U.S. Pat. No. 11,694,702 issued Jul. 4, 2023), which is a continuation of U.S. application Ser. No. 16/527,160 filed Jul. 31, 2019 (now U.S. Pat. No. 11,120,809 issued Sep. 14, 2021), which is a continuation of U.S. application Ser. No. 15/307,059 filed Oct. 27, 2016 (now U.S. Pat. No. 10,418,042 issued Sep. 17, 2019), the entire contents of which are incorporated herein by reference. U.S. application Ser. No. 15/307,059 is a National Stage of PCT/JP2015/057727 filed Mar. 16, 2015, which claims the benefit of priority under 35 U.S.C. §119 from Japanese Application No. 2014-094758 filed May 1, 2014.

The present invention relates to a coding technology and a decoding technology of coding and decoding linear prediction coefficients and coefficients which are convertible thereinto.

In coding of sound signals such as speech and music, a method of performing the coding by using linear prediction coefficients obtained by performing linear prediction analysis on an input sound signal is widely used.

In order to make it possible to obtain, on the part of a decoding device, the information on the linear prediction coefficients used in coding processing by decoding, a coding device codes the linear prediction coefficients and sends a code corresponding to the linear prediction coefficients to the decoding device. In Non-patent Literature 1, a coding device converts linear prediction coefficients into a sequence of LSP (Line Spectrum Pair) parameters which are parameters in a frequency domain and equivalent to the linear prediction coefficients and sends an LSP code obtained by coding the sequence of LSP parameters to a decoding device.

In Non-patent Literature 1, in order to reduce the code amount of the LSP code, a vector coding and decoding technology using moving average prediction (MA prediction) is used.

First, the flow of coding processing will be described.

depicts the configuration of an existing linear prediction coefficient coding device.

To the linear prediction coefficient coding device, LSP (Line Spectrum Pairs) parameters θ[1], θ[2], . . . , θ[p] of each frame are input, and the linear prediction coefficient coding deviceperforms the following processing of a predictive subtraction unit, a vector coding unit, and a delay input uniton a frame-by-frame basis, obtains an LSP code C, and outputs the LSP code C. Incidentally, f represents a frame number and p represents a prediction order.

When an input sound signal Xis input to the linear prediction coefficient coding device, the linear prediction coefficient coding deviceis also provided with a linear prediction analysis unitand an LSP calculation unit, and the frame-by-frame input sound signals Xare consecutively input thereto and the following processing is performed on a frame-by-frame basis.

Hereinafter, specific processing of each unit will be described.

The linear prediction analysis unitreceives the input sound signal X, performs linear prediction analysis on the input sound signal X, obtains linear prediction coefficients a[1], a[2], . . . , a[p], and outputs the linear prediction coefficients a[1], a[2], . . . , a[p]. Here, a[i] represents an ith-order linear prediction coefficient that is obtained by performing linear prediction analysis on an input sound signal Xof an fth frame.

The LSP calculation unitreceives the linear prediction coefficients a[1], a[2], . . . , a[p], obtains LSP parameters θ[1], θ[2], . . . , θ[p] from the linear prediction coefficients a[1], a[2], . . . , a[p], and outputs an LSP parameter vector Θ=(θ[1], θ[2], . . . , θ[p])that is a vector using the obtained LSP parameters as elements thereof. Here, θ[i] is an ith-order LSP parameter corresponding to the input sound signal Xof the fth frame.

The predictive subtraction unitis formed of, for example, a storagestoring a predetermined coefficient α, a storagestoring a predictive mean vector V, a multiplication unit, and subtraction unitsand

The predictive subtraction unitreceives the LSP parameter vector Θand a preceding-frame quantization differential vector {circumflex over ( )}S.

The predictive subtraction unitgenerates a differential vector S=Θ−V−α×{circumflex over ( )}S=(s[1], s[2], . . . , s[p])that is a vector obtained by subtracting the predictive mean vector V and a vector α{circumflex over ( )}Sfrom the LSP parameter vector Θand outputs the differential vector S.

Incidentally, the predictive mean vector V=(v[1], v[2], . . . , v[p])is a predetermined vector stored in the storageand simply has to be obtained in advance from, for example, a sound signal for learning. For example, in the linear prediction coefficient coding device, by using a sound signal picked up in the same environment (for instance, the same speaker, sound pick-up device, and place) as the sound signal to be coded as an input sound signal for learning, LSP parameter vectors of many frames are obtained, and the average thereof is used as the predictive mean vector.

The multiplication unitobtains a vector α×{circumflex over ( )}Sby multiplying a decoded differential vector {circumflex over ( )}Sof a preceding frame by the predetermined coefficient α stored in the storage

Incidentally, in, by using the two subtraction unitsand, first, after the predictive mean vector V stored in the storageis subtracted from the LSP parameter vector Θin the subtraction unit, the vector α×{circumflex over ( )}Sis subtracted in the subtraction unit, but the above may be performed the other way around. Alternatively, the differential vector Smay be generated by subtracting, from the LSP parameter vector Θ, a vector V+α×{circumflex over ( )}Sobtained by adding the predictive mean vector V and the vector α×{circumflex over ( )}S.

The differential vector Sof the present frame may also be called a vector that is obtained by subtracting a vector containing at least a prediction based on a past frame from a vector (an LSP parameter vector Θ) based on coefficients which are convertible into linear prediction coefficients of more than one order of the present frame.

The vector coding unitreceives the differential vector S, codes the differential vector S, and obtains an LSP code Cand a quantization differential vector {circumflex over ( )}S=({circumflex over ( )}s[1], {circumflex over ( )}s[2], . . . , {circumflex over ( )}s[p])corresponding to the LSP code Cand outputs the LSP code Cand the quantization differential vector {circumflex over ( )}S. For coding of the differential vector S, any one of the well-known coding methods may be used, such as a method of vector quantizing the differential vector S, a method of dividing the differential vector Sinto a plurality of subvectors and vector quantizing each of the subvectors, a method of multistage vector quantizing the differential vector Sor the subvectors, a method of scalar quantizing the elements of a vector, and a method obtained by combining these methods.

Here, an example of a case in which the method of vector quantizing the differential vector Sis used will be described.

The vector coding unitsearches for a candidate differential vector closest to the differential vector Sfrom a plurality of candidate differential vectors stored in a vector codebookand outputs the candidate differential vector as the quantization differential vector {circumflex over ( )}S, and outputs a differential vector code corresponding to the quantization differential vector {circumflex over ( )}Sas the LSP code C. Incidentally, the quantization differential vector {circumflex over ( )}Scorresponds to a decoded differential vector which will be described later.

In the vector codebook, candidate differential vectors and differential vector codes corresponding to the candidate differential vectors are stored in advance.

The delay input unitreceives the quantization differential vector {circumflex over ( )}S, holds the quantization differential vector {circumflex over ( )}S, delays the quantization differential vector {circumflex over ( )}Sby one frame, and outputs the resultant vector as a preceding-frame quantization differential vector {circumflex over ( )}S. That is, if the predictive subtraction unithas performed processing on a quantization differential vector {circumflex over ( )}Sof an fth frame, the delay input unitoutputs a quantization differential vector {circumflex over ( )}Son an f−1th frame.

depicts the configuration of an existing linear prediction coefficient decoding device. To the linear prediction coefficient decoding device, frame-by-frame LSP codes Care consecutively input, and the linear prediction coefficient decoding deviceobtains a decoded predictive LSP parameter vector {circumflex over ( )}Θ=({circumflex over ( )}θ[1], {circumflex over ( )}θ[2], . . . , {circumflex over ( )}θ[p]) by decoding the LSP code Con a frame-by-frame basis.

Hereinafter, specific processing of each unit will be described.

A vector decoding unitreceives the LSP code C, decodes the LSP code C, obtains a decoded differential vector {circumflex over ( )}Scorresponding to the LSP code C, and outputs the decoded differential vector {circumflex over ( )}S. For decoding of the LSP code C, a decoding method corresponding to the coding method adopted by the vector coding unitof the coding device is used.

Here, an example of a case in which a decoding method corresponding to the method adopted by the vector coding unit, the method of vector quantizing the differential vector S, is used will be described.

The vector decoding unitsearches for a plurality of differential vector codes corresponding to the LSP code Cfrom differential vector codes stored in a vector codebookand outputs a candidate differential vector corresponding to the differential vector codes as the decoded differential vector {circumflex over ( )}S. Incidentally, the decoded differential vector {circumflex over ( )}Scorresponds to the above-described quantization differential vector {circumflex over ( )}Sand corresponding elements take the same values if there are no transmission errors and no errors and the like in the course of coding and decoding.

In the vector codebook, the candidate differential vectors and the differential vector codes corresponding to the candidate differential vectors are stored in advance. Incidentally, the vector codebookshares information in common with the vector codebookof the above-described linear prediction coefficient coding device.

A delay input unitreceives the decoded differential vector {circumflex over ( )}S, holds the decoded differential vector {circumflex over ( )}S, delays the decoded differential vector {circumflex over ( )}Sby one frame, and outputs the resultant vector as a preceding-frame decoded differential vector {circumflex over ( )}S. That is, if a predictive addition unitperforms processing on a decoded differential vector {circumflex over ( )}Sof an fth frame, the delay input unitoutputs a decoded differential vector {circumflex over ( )}Sof an f−1th frame.

A predictive addition unitis formed of, for example, a storagestoring a predetermined coefficient α, a storagestoring a predictive mean vector V, a multiplication unit, and addition unitsand

The predictive addition unitreceives the decoded differential vector {circumflex over ( )}Sof the present frame and the preceding-frame decoded differential vector {circumflex over ( )}S.

The predictive addition unitgenerates a decoded predictive LSP parameter vector {circumflex over ( )}Θ(={circumflex over ( )}S+V+α{circumflex over ( )}S) that is a vector obtained by adding the decoded differential vector {circumflex over ( )}S, the predictive mean vector V=(v[1], v[2], . . . , v[N]), and a vector α×{circumflex over ( )}Sand outputs the decoded predictive LSP parameter vector {circumflex over ( )}Θ.

The multiplication unitobtains the vector α×{circumflex over ( )}Sby multiplying the preceding-frame decoded differential vector {circumflex over ( )}Sby the predetermined coefficient α stored in the storage

In, by using the two addition unitsand, first, after the vector α×{circumflex over ( )}Sis added to the decoded differential vector {circumflex over ( )}Sof the present frame in the addition unit, the predictive mean vector V is added in the addition unit, but the above may be performed the other way around. Alternatively, the decoded predictive LSP parameter vector {circumflex over ( )}Θmay be generated by adding a vector obtained by adding the vector α×{circumflex over ( )}Sand the predictive mean vector V to the decoded differential vector {circumflex over ( )}S.

Incidentally, it is assumed that the predictive mean vector V used here is the same as the predictive mean vector V used in the predictive subtraction unitof the above-described linear prediction coefficient coding device.

If linear prediction coefficients are necessary, the linear prediction coefficient decoding devicemay be provided with a decoded predictive linear prediction coefficient calculation unit. In this case, the decoded predictive linear prediction coefficient calculation unitreceives the decoded predictive LSP parameter vector {circumflex over ( )}Θ, converts the decoded predictive LSP parameter vector {circumflex over ( )}Θinto decoded predictive linear prediction coefficients {circumflex over ( )}a[1], {circumflex over ( )}a[2], . . . , {circumflex over ( )}a[p], and outputs the decoded predictive linear prediction coefficients {circumflex over ( )}a[1], {circumflex over ( )}a[2], . . . , {circumflex over ( )}a[p].

Non-patent Literature 1: “ITU-T Recommendation G.729”, ITU, 1996

In the linear prediction coefficient decoding device of Non-patent Literature 1, since decoding processing of an fth frame is performed by using the decoded differential vector {circumflex over ( )}Swhich is the decoding result of an f−1th frame, not only when a transmission error occurs in an LSP code of the present frame, but also when a transmission error occurs in an LSP code of the immediately preceding frame, LSP parameters of the present frame also cannot be decoded correctly.

In the linear prediction coefficient decoding device of Non-patent Literature 1, since the LSP parameters obtained by decoding are used only for linear prediction synthesis, even when the LSP parameters cannot be decoded correctly, this merely causes a reduction in the sound quality of the decoded sound signal in a plurality of consecutive frames. That is, it can be said that the linear prediction coefficient coding device and the linear prediction coefficient decoding device of Non-patent Literature 1 have a configuration which gives a higher priority to expressing the LSP parameters with a small code amount than to a problem which will arise when the LSP parameters cannot be decoded correctly.

However, the linear prediction coefficient coding device and the linear prediction coefficient decoding device are also used in a coding device and a decoding device which use the LSP parameters not only for linear prediction analysis and synthesis, but also for variable-length coding and decoding depending on the amplitude values forming a spectral envelope which is determined from the LSP parameters. In this case, the following problem arises: if the LSP parameters cannot be decoded correctly in one frame, variable-length decoding cannot be performed correctly in a plurality of consecutive frames including that frame, which makes it impossible to obtain a decoded sound signal.

In view of such a problem, an object of the present invention is to provide a coding method and a decoding method of coding and decoding coefficients which are convertible into linear prediction coefficients, the coding method and the decoding method that can use in combination predictive coding method and decoding method which are a coding method and a decoding method that can accurately express coefficients which are convertible into linear prediction coefficients with a small code amount, the coefficients such as those used in linear prediction analysis and synthesis, for example, and a coding method and a decoding method which can obtain correctly, by decoding, coefficients which are convertible into linear prediction coefficients of the present frame, even when a linear prediction coefficient code (for example, an LSP code) that is a code corresponding to coefficients which are convertible into linear prediction coefficients of a preceding frame, the coefficients such as those used in variable-length coding/decoding depending on the amplitude values forming a spectral envelope which is determined from LSP parameters, for example, is not correctly input to a linear prediction coefficient decoding device, if a linear prediction coefficient code of the present frame is correctly input to the linear prediction coefficient decoding device.

In order to solve the above-described problem, according to one aspect of the present invention, a coding device includes: a predictive coding unit that obtains a first code by coding a differential vector formed of differentials between a vector of coefficients which are convertible into linear prediction coefficients of more than one order of a present frame and a prediction vector containing at least a predicted vector from a past frame, and obtains a quantization differential vector corresponding to the first code; and a non-predictive coding unit that generates a second code by coding a correction vector which is formed of differentials between the vector of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame and the quantization differential vector or formed of some of elements of the differentials.

In order to solve the above-described problem, according to another aspect of the present invention, a coding device includes: a predictive coding unit that obtains a first code by coding a differential vector formed of differentials between a vector of coefficients which are convertible into linear prediction coefficients of more than one order of a present frame and a prediction vector formed of at least a prediction based on a past frame and a predetermined vector, and obtains a quantization differential vector corresponding to the first code; and a non-predictive coding unit that generates a second code by coding a correction vector which is formed of differentials obtained by subtracting the quantization differential vector and the predetermined vector from the vector of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame or formed of some of elements of the differentials.

In order to solve the above-described problem, according to another aspect of the present invention, a decoding device includes: a predictive decoding unit that obtains a decoded differential vector by decoding a first code and generates a first decoded vector formed of decoded values of coefficients which are convertible into linear prediction coefficients of more than one order of a present frame by adding the decoded differential vector and a prediction vector containing at least a prediction based on a past frame; and a non-predictive decoding unit that obtains a decoded correction vector by decoding a second code and generates a second decoded vector formed of decoded values of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame by adding elements of the decoded correction vector and at least elements of corresponding orders of the decoded differential vector.

In order to solve the above-described problem, according to another aspect of the present invention, a decoding device includes: a predictive decoding unit that obtains a decoded differential vector by decoding a first code and generates a first decoded vector formed of decoded values of coefficients which are convertible into linear prediction coefficients of more than one order of a present frame by adding the decoded differential vector and a prediction vector formed of at least a prediction based on a past frame and a predetermined vector; and a non- predictive decoding unit that obtains a decoded correction vector by decoding a second code and generates a second decoded vector formed of decoded values of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame by adding, to the decoded correction vector, at least the decoded differential vector and the predetermined vector for each of elements of corresponding orders.

In order to solve the above-described problem, according to another aspect of the present invention, a coding method includes: a predictive coding step of obtaining a first code by coding a differential vector formed of differentials between a vector of coefficients which are convertible into linear prediction coefficients of more than one order of a present frame and a prediction vector containing at least a predicted vector from a past frame, and obtaining a quantization differential vector corresponding to the first code; and a non-predictive coding step of generating a second code by coding a correction vector which is formed of differentials between the vector of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame and the quantization differential vector or formed of some of elements of the differentials.

In order to solve the above-described problem, according to another aspect of the present invention, a coding method includes: a predictive coding step of obtaining a first code by coding a differential vector formed of differentials between a vector of coefficients which are convertible into linear prediction coefficients of more than one order of a present frame and a prediction vector formed of at least a prediction based on a past frame and a predetermined vector, and obtaining a quantization differential vector corresponding to the first code; and a non-predictive coding step of generating a second code by coding a correction vector which is formed of differentials obtained by subtracting the quantization differential vector and the predetermined vector from the vector of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame or formed of some of elements of the differentials.

In order to solve the above-described problem, according to another aspect of the present invention, a decoding method includes: a predictive decoding step of obtaining a decoded differential vector by decoding a first code and generating a first decoded vector formed of decoded values of coefficients which are convertible into linear prediction coefficients of more than one order of a present frame by adding the decoded differential vector and a prediction vector containing at least a prediction based on a past frame; and a non-predictive decoding step of obtaining a decoded correction vector by decoding a second code and generating a second decoded vector formed of decoded values of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame by adding elements of the decoded correction vector and at least elements of corresponding orders of the decoded differential vector.